Comparing Open Source Digital Library Software [chapter]

George Pyrounakis, Mara Nikolaidou
Handbook of Research on Digital Libraries  
Greenstone: Indexing is offered for the text documents and specific metadata fields. Searching capabilities provided for defined sections in a document (Title, chapter, paragraph) or in whole document. Stemming and case sensitive searching is also available. Managing Gigabytes (MG) open-source applications is used to support indexing and searching. Browsing catalogs can be defined for specific fields using hierarchical structure. Keystone: Indexing is supported on specified document types for
more » ... e whole metadata set. Free text searching is offered. Browsing mechanism is not provided. EPrints: Indexing is supported for every metadata field, using the MySQL database. Full text indexing is supported for selected fields. Combined fielded search and free text search are provided to the end-user. Browsing is provided using specified fields (e.g. title, author, subject). Object management DSpace: Items in DSpace are created using the web submission user interface or the batch item importer, which ingests XML metadata documents and the constituent content files. In both cases a workflow process may initiate depending on the collection configuration. The workflow can be configured to contain from one to three steps where different users or groups may intervene to the item submission. Collections and communities are created using the web user interface. Fedora: Creation of digital objects is feasible using the Administrator client or the batch import utility (XML files in METS or FOXML format). Metadata addition or editing is provided through a text editor in Administrator client. The same client is used for addition and removal of digital content (as datastreams). Greenstone: New collections and the contained documents are built using the Greenstone Librarian Interface or the command line building program. Keystone: The content management system of Keystone provides the web interface for editing documents. It allows specified users to manage the content of documents as long as the files structure. EPrints: A default web user interface is provided for the creation and editing of objects. Authority records can be used helping the completion of specific fields (e.g. authors, title). Objects can also be imported from text files using multiple formats (METS, DC, MODS, BibTeX, EndNote). User interfaces DSpace: A default web user interface is provided in order for the end-user to browse a collection, view the qualified DC metadata of an item and navigate to its bistreams. Navigation into an item is supported through the structural metadata that may determine the ordering of complex content (like book pages or web pages). A searching interface is provided by default that allows the user to search using keywords. Fedora: The web interface of Fedora provides a search environment to the end-user, where he/she may execute simple keyword or field search queries. The default view of digital objects is restricted to the presentation of the system metadata and the datastreams. Behavior digital objects define the presentation or manipulation methods of datastreams. A developer may build specific web services and attach them on digital objects as behaviors. A DC metadata viewing page and an image manipulation applet are provided as default behaviors. Greenstone: The default web user interface provides browsing and searching into collections, navigating into hierarchical objects (like books) using table of contents. Presentation of documents or search results may differ depending on specified XSLTs. Keystone: Presentation of a document is controlled by an XSLT stylesheet that reflects the associated document type. The main web user interface is based on a portal like environment. In this environment a user may browse the documents directory structure and search in the digital library. EPrints: The web user interface provides browsing by selected metadata fields (usually subject, title or date). Browsing can be hierarchical for subject fields. Searching environment allows user to restrict the search query using multiple fields and select values from lists. Access control DSpace: It supports users (e-people) and groups that hold different rights. Authentication is provided through user passwords, X509 certificates or LDAP. Access control rights are kept for each item and define the actions that a user is able to perform. These actions are: read/write the bitstreams of an item, add/remove the bundles of an item, read/write an item, add/remove an item in a collection. Rights are based in a default-deny policy. Fedora: It supports users and groups authorized for accessing specific digital objects using XACML policies. Authentication is provided through LDAP or for specific IP addresses. Greenstone: A user in Greenstone belongs to one of two predefined user groups: an administrator or a collection builder. The first user group has the right to create and delete users, while the second builds and updates collections. End-users have access to all the collections and the documents. Keystone: A simple access control is supported where you can define administrators and simple users that have access rights on specific parts of the documents structure. EPrints: Registered users in EPrints are able to create and edit objects. Users are logged in using their username and password pair. Multiple languages support All the DL systems use Unicode character encoding, so the support of different languages can be supported. Every system can use multiple languages in the metadata fields and digital content. Keystone and EPrints provide an XML attribute on metadata fields to define the language used for the field value. Greenstone provides ready to use multilingual interfaces already translated in many languages. Interoperability features All the DL systems support OAI-PMH in order to share the metadata of the DL with other repositories. Greenstone and Keystone also support Z39.50 protocol for answering queries on specific metadata sets. Fedora and DSpace are able to export digital objects as METS XML files. Both systems also use persistence URIs to access the digital content providing a unified access mechanism to external services. DSpace also supports OpenURL protocol providing links for every item page. EPrints exports data objects in METS and MPEG-21 Digital Item Declaration Language (DIDL) format. Level of customization Dspace: Although DSpace has a flexible object model is not so open in constructing very different objects with independent metadata sets because of its database oriented architecture. The user interface is fixed and provides only minor presentation interventions. Another disadvantage is the full support of only specific file formats as digital content. Fedora: In Fedora every digital object can follow a different content model that describes its format. It is also possible to provide multiple behaviors in it that determine the access and manipulation methods of the digital object. These two characteristics result in a fully customizable DL. The user interface, although by default is poor, is fully customizable based on two APIs (Access API and Management API). Greenstone: It provides customization for the presentation of a collection based on XSLTs and agents that control specific actions of the DL. Greenstone architecture provides (i) a back end that contains the collections and the documents as long as services to manage them and (ii) a web based front end that is responsible for the presentation of collections, documents and their searching environment. Keystone: Document's structure is based on a customized document type, which is formed by an XML Schema. In addition the presentation of a document is dependent on the XSLTs associated with the document type. The separation of document storage and presentation layer, as long as the typing of documents provides a fully customizable DL architecture. EPrints: The data objects in EPrints contain user defined metadata. Plug-ins can be written in order to export the data objects in different text formats. A Core API in Perl is provided for developers who prefer to access basic DL functionality. Based on the above analysis, the five DL systems were graded for each of the characteristics. The minimum score is 1 and the maximum is 5.
doi:10.4018/978-1-59904-879-6.ch006 fatcat:pkjccfofejew5mzf3zpr75xypi